On Mon, Mar 20, 2023 at 3:37 AM Peter Smith <smithpb2...@gmail.com> wrote: > > There are a couple of TAP tests where the copy binary is expected to > fail. And when it fails, you do binary=false (changing the format back > to 'text') so the test is then expected to be able to proceed. > > I don't know if this happens in practice, but IIUC in theory, if the > timing is extremely bad, the tablesync could relaunch in binary mode > multiple times (any fail multiple times?) before your binary=false > change takes effect. > > So, I was wondering if it could help to use the subscription > 'disable_on_error=true' parameter for those cases so that the > tablesync won't needlessly attempt to relaunch until you have set > binary=false and then re-enabled the subscription. >
+1. That would make tests more reliable. -- With Regards, Amit Kapila.